Radial Basis Function for Breast Lesion Detection from MammoWave Clinical Data

Abstract

Recently, a novel microwave apparatus for breast lesion detection (MammoWave), uniquely able to function in air with 2 antennas rotating in the azimuth plane and operating within the band 1–9 GHz has been developed. Machine learning (ML) has been implemented to understand information from the frequency spectrum collected through MammoWave in response to the stimulus, segregating breasts with and without lesions. The study comprises 61 breasts (from 35 patients), each one with the correspondent output of the radiologist’s conclusion (i.e., gold standard) obtained from echography and/or mammography and/or MRI, plus pathology or 1-year clinical follow-up when required. The MammoWave examinations are performed, recording the frequency spectrum, where the magnitudes show substantial discrepancy and reveals dissimilar behaviours when reflected from tissues with/without lesions. Principal component analysis is implemented to extract the unique quantitative response from the frequency response for automated breast lesion identification, engaging the support vector machine (SVM) with a radial basis function kernel. In-vivo feasibility validation (now ended) of MammoWave was approved in 2015 by the Ethical Committee of Umbria, Italy (N. 6845/15/AV/DM of 14 October 2015, N. 10352/17/NCAV of 16 March 2017, N 13203/18/NCAV of 17 April 2018). Here, we used a set of 35 patients. According to the radiologists conclusions, 25 breasts without lesions and 36 breasts with lesions underwent a MammoWave examination. The proposed SVM model achieved the accuracy, sensitivity, and specificity of 91%, 84.40%, and 97.20%. The proposed ML augmented MammoWave can identify breast lesions with high accuracy.

Review and Analysis of Tumour Detection and Image Quality Analysis in Experimental Breast Microwave Sensing

“…Table 3 summarizes these investigations into ML-based BMS cancer diagnosis. Three additional articles are noteworthy [80,104,113], but are excluded from Table 3. The work in [104] used the same dataset as [118] and used tree-based models, but none of the models improved upon the results in [118].…”

“…The work in [104] used the same dataset as [118] and used tree-based models, but none of the models improved upon the results in [118]. The methods and results in [113] are expanded upon in [202], and therefore only [202] is included in Table 3. The authors of [80] used a dataset that combined experimental and simulated data.…”

“…Notably, the MammoWave system [135] has been used in five ML-based investigations [112,113,135,188,202]. These works provide the most optimistic estimates of the diagnostic performance of BMS in the literature, but the reported results are likely overestimates due to methodological limitations in these works.…”

Gradient-Boosting Algorithm for Microwave Breast Lesion Classification—SAFE Clinical Investigation

“…Recently, UBT published its work considering automated breast lesion identification. A machine learning (ML) model, namely the support vector machine (SVM), was implemented on the raw data of 35 patients, achieving accuracy, sensitivity, and specificity of 91%, 84.4% and 97.2%, respectively [ 34 ]. Mitos Medical Technologies (MITOS), Turkey, in collaboration with Istanbul Technical University, Turkey, published the results of their clinical research study conducted with their SAFE (Scan and Find Early) MBI device, where the reported sensitivity was 63% [ 35 ].…”

Special Issue “Advancement in Breast Diagnostic and Interventional Radiology”

“…Experimental models of new methods such as MammoWave are one of the new efforts to improve breast imaging. It is very difficult to identify an option that can support methods with high diagnostic accuracy, However, experiments with results that are negative or not positive, as well as those that end in failure, are part of the development and allow research to continue [ 11 ].…”